α- And β-subunit composition of voltage-gated sodium channels investigated with μ-conotoxins and the recently discovered μO§-conotoxin GVIIJ

Abstract

We investigated the identities of the isoforms of the α (Na_V1)- and β (Na_Vβ)-subunits of voltage-gated sodium channels, including those responsible for action potentials in rodent sciatic nerves. To examine α-subunits, we used seven μ-conotoxins, which target site 1 of the channel. With the use of exogenously expressed channels, we show that two of the μ-conotoxins, μ-BuIIIB and μ-SxIIIA, are 50-fold more potent in blocking Na_V1.6 from mouse than that from rat. Furthermore, we observed that μ-BuIIIB and μ-SxIIIA are potent blockers of large, myelinated A-fiber compound action potentials (A-CAPs) [but not small, unmyelinated C-fiber CAPs (C-CAPs)] in the sciatic nerve of the mouse (unlike A-CAPs of the rat, previously shown to be insensitive to these toxins). To investigate β-subunits, we used two synthetic derivatives of the recently discovered μO§-conotoxin GVIIJ that define site 8 of the channel, as previously characterized with cloned rat Na_V1- and Na_Vβ-subunits expressed in Xenopus laevis oocytes, where it was shown that μO§-GVIIJ is a potent inhibitor of several Na_V1-isoforms and that coexpression of Na_Vβ2 or -β4 (but not Na_Vβ1 or -β3) totally protects against block by μO§-GVIIJ. We report here the effects of μO§-GVIIJ on 1) sodium currents of mouse Na_V1.6 coexpressed with various combinations of Na_Vβ-subunits in oocytes; 2) A- and C-CAPs of mouse and rat sciatic nerves; and 3) sodium currents of small and large neurons dissociated from rat dorsal root ganglia. Our overall results lead us to conclude that action potentials in A-fibers of the rodent sciatic nerve are mediated primarily by Na_V1.6 associated with Na_Vβ2 or Na_Vβ4.

voltage-gated sodium channels (VGSCs) are responsible for the upshoot of action potentials. Each VGSC consists of an α-subunit (of which there are nine isoforms, Na_V1.1–Na_V1.9) and one or more Na_Vβ-subunits (of which there are four isoforms, Na_Vβ1–Na_Vβ4), likely either an Na_Vβ1- or Na_Vβ3-subunit and either an Na_Vβ2- or Na_Vβ4-subunit to form a ternary, heterotrimeric αββ complex (Calhoun and Isom 2014; Catterall 2012). The α-subunit has some 2,000 amino acid (AA) residues and consists of four homologous domains (DI–DIV), each with six membrane-spanning segments (S1–S6). Each domain has a voltage-sensor module (comprised of S1–S4) and a pore module (comprised of S5–S6, with their connecting re-entrant loop consisting of short segments SS5 and SS6), and the four domains are radially arranged, with each pore module contributing to a central, ion-conducting pore (Catterall 2014). The β-subunit has some 200 AA residues and forms a single transmembrane-spanning segment that separates a large, extracellular domain from a smaller, intracellular C-terminal tail. The Na_Vβ1- or Na_Vβ3-subunit is noncovalently associated with the α-subunit, whereas the Na_Vβ2- or Na_Vβ4-subunit is covalently coupled to the α-subunit via a disulfide bond located on the extracellular aspect of the subunits (Calhoun and Isom 2014; Catterall 2012; Chen et al. 2012; Gilchrist et al. 2013; Yu et al. 2003). A given neuron can have more than one single isoform of an α-subunit, and we are developing approaches to identify the functional contributions of each isoform through the use of conotoxins (Wilson et al. 2011a; Zhang et al. 2013b).

Five families of conotoxins that target sodium channels have been identified thus far. Three families consist of peptides that are antagonists: μ-conotoxins, which are pore blockers, like TTX, and compete with TTX in binding to site 1 (Cestèle and Catterall 2000; Cruz et al. 1985; Zhang et al. 2009, 2010a); μO-conotoxins, which are gating modifiers that inhibit channel activation by interacting with site 4, the extracellular loop connecting S3 and S4 of DII (Heinemann and Leipold 2007; Leipold et al. 2007); and μO§-conotoxins, which bind to site 8, centered on a Cys residue between S5 and SS5 of DII [specifically, C910 in the case of rat Nav1.2 (rNa_V1.2)] but whose mechanism of block remains to be established (Gajewiak et al. 2014). The other two conotoxin families consist of peptides that are VGSC agonists: δ-conotoxins, which inhibit channel inactivation by binding to site 6 at the extracellular loop between S3 and S4 of DIV (Leipold et al. 2005); and ι-conotoxins, which promote channel activation but whose site of action on VGSCs remains to be established (Fiedler et al. 2008). [For a recent review, see Stevens et al. (2011).] These toxins possess varying degrees of Na_V1 specificities when examined with Xenopus laevis oocytes expressing cloned Na_V1s.

A compound action potential (CAP) consists of extracellularly recorded action potentials of a population of axons (or fibers). A- and C-CAPs are conducted by large, myelinated A-fibers and small, unmyelinated C-fibers, respectively; thus A-CAPs have faster conduction velocities than C-CAPs, and their waveforms are readily distinguishable by their latencies. We used a panel of μ-conotoxins to conclude that A- and C-CAPs in the rat sciatic nerve are mediated principally by Na_V1.6 and -1.7, respectively (Wilson et al. 2011a). We also used μ-conotoxins to show that Na_V1.1, -1.6, and -1.7 could account for all of the TTX-sensitive voltage-gated sodium currents (I_Na) in cell bodies of acutely dissociated rat dorsal root ganglia (DRG) neurons; furthermore, the levels of functional Na_V1s in large neurons were Na_V1.7 ≤ Na_V1.1 ≤ Na_V1.6. In a class of small neurons, whose I_Na was >50% sensitive to TTX, the levels were Na_V1.1 ≤ Na_V1.6 ≪ Na_V1.7 (Zhang et al. 2013b).

In contrast to Na_V1.6 from mouse (mNa_V1.6), much less has been reported concerning the functional and pharmacological properties of Na_V1.6 from rat (rNa_V1.6) since its initial description (Dietrich et al. 1998) until relatively recently (Gajewiak et al. 2014; He and Soderlund 2014; Tan et al. 2011; Tan and Soderlund 2011; Zhang et al. 2013a). While comparing the abilities of various μ-conotoxins in blocking mNa_V1.6 expressed in oocytes and A-fibers of the rat sciatic nerve, we encountered an apparent discrepancy; that is, at the time of those experiments, only an mNa_V1.6 clone was available to us, and we noted that two of the μ-conopeptides tested, μ-SxIIIA and μ-BuIIIB, which were potent in blocking mNa_V1.6 expressed in oocytes, were unable to block A-CAPs in the rat sciatic nerve, unlike five other μ-conotoxins (μ-SmIIIA, μ-KIIIA, μ-SIIIA, μ-GIIIA, and μ-PIIIA) that readily blocked both rat A-CAPs and mNa_V1.6 (Wilson et al. 2011a). [It might be noted that all five of the latter μ-conotoxins blocked mNa_V1.6 with submicromolar affinities (Wilson et al. 2011a), and two of them, μ-SmIIIA and μ-PIIIA, were tested on rNa_V1.6, where they also blocked with submicromolar affinities (Zhang et al. 2013a).] To reconcile the discrepancy, we suggested that the two μ-conotoxins, μ-SxIIIA and μ-BuIIIB, might be better able to block mNa_V1.6 than rNa_V1.6 (Wilson et al. 2011a). With the use of our recently acquired clone of rNa_V1.6 (Zhang et al. 2013a), we show here that this is indeed the case.

The affinities of μ-conotoxins for rNa_V1s can be affected by coexpression of Na_Vβ-subunits (Zhang et al. 2013a); thus we examined how the affinity of μ-SxIIIA for mNa_V1.6 is affected by coexpression with various combinations of rNa_Vβ-subunits in oocytes (note, only rat clones of Na_Vβ-subunits are presently available to us). Additionally, we tested μ-SxIIIA and μ-BuIIIB along with five other μ-conotoxins on A- and C-CAPs of the mouse sciatic nerve to help identify likely Na_V1-isoforms responsible for the propagation of action potentials in mouse A- and C-fibers.

We also tested two closely related synthetic derivatives of the recently discovered μO§-conotoxin GVIIJ, namely, μO§-GVIIJ_SSC and μO§-GVIIJ_SSG, where the subscripts refer, respectively, to cysteinylated and glutathionylated analogs that share the same peptide backbone (see materials and methods). μO§-GVIIJ_SSG was previously tested on oocytes expressing rNa_V1.1–rNa_V1.8, and it potently blocked all except Na_V1.5 and -1.8, which were blocked poorly and not at all, respectively (Gajewiak et al. 2014). Except for a small (approximately threefold) difference in association rate constant (k_on) (Gajewiak et al. 2014), both derivatives behave similarly in all functional tests performed thus far, including those reported here. For historical reasons, the early experiments were largely performed with μO§-GVIIJ_SSG, whereas more recent experiments involved μO§-GVIIJ_SSC, because its structure turned out to resemble more closely that of the native peptide (Gajewiak et al. 2014) (see materials and methods). For brevity, the condensed term μO§-GVIIJ_SSC/G will be used when referring to both peptides.

The binding site of μO§-GVIIJ_SSC/G on the channel, site 8, is spatially distinct from those of site 1 (where μ-conotoxins bind) and site 4 (where μO-conotoxins bind), insofar as electrophysiological tests reveal that neither the μ-conotoxin derivative μ-KIIIA[K7A] nor μO-conotoxin MrVIB interferes with the block of rNa_V1.2 by μO§-GVIIJ_SSG (Gajewiak et al. 2014).

A novel feature of μO§-GVIIJ_SSC/G, originally observed with μO§-GVIIJ_SSG, is that coexpression in oocytes of rNa_V1s with either Na_Vβ2 or -β4 protects the channel against block by the peptide (Gajewiak et al. 2014). We demonstrate here that this is also true for mNa_V1.6; specifically, coexpression with Na_Vβ2 or -β4 (but not Na_Vβ1 or -β3) protects the channel against block by μO§-GVIIJ_SSC. We also show that unlike members of all other conopeptide families that target VGSCs, μO§-GVIIJ_SSC/G had no effect on A- and C-CAPs of rat and mouse sciatic nerves. Finally, we examined the ability of μO§-GVIIJ_SSG to block I_Na, mediated by VGSCs endogenously expressed in acutely dissociated neurons of rat DRG, and observed that I_Na of small, but not large, neurons could be blocked by the peptide.

MATERIALS AND METHODS

Toxins.

μ-Conotoxins were synthesized as described previously (Wilson et al. 2011a). TTX was obtained from Alomone Labs (Jerusalem, Israel). δ-Conotoxin PVIA (δ-PVIA) was synthesized as described previously (Bulaj et al. 2001).

The sequence of μO§-GVIIJ is as follows: GWCGDOGATCGKLRLYCCSGFCDCYTKTCKDKSSÂ, where the seven Cys residues are in boldface, O represents hydroxyproline, the underlined C is Cys24, and the caret signifies a free carboxyl terminus. In the native peptide, Trp2 is bromoTrp, and Cys24 is disulfide bonded to a cysteine; i.e., Cys24 is S-cysteinylated (Gajewiak et al. 2014). Two derivatives of the peptide, μO§-GVIIJ_SSC and μO§-GVIIJ_SSG, were synthesized as recently described (Gajewiak et al. 2014) and used in the present study. Trp2 in neither derivative was brominated, and the two derivatives differed from each other in that Cys24 was disulfide bonded either to cysteine (in μO§-GVIIJ_SSC) or glutathione (in μO§-GVIIJ_SSG). (Thus μO§-GVIIJ_SSC differs from the native peptide by only lacking bromination of Trp2.) The two peptides sharing the condensed term, μO§-GVIIJ_SSC/G, behaved similarly in blocking rNa_V1s expressed in oocytes (Gajewiak et al. 2014). Their use in a given experiment of the present study was based, in part, on their availability at the time the experiment was performed.

Preparation and voltage clamp of X. laevis oocytes expressing cloned VGSCs.

Clones for mNa_v1.6 (NM_011323), rNa_Vβ1 (NM_017288), and rNa_Vβ2 (NM_012877.1) were obtained from Alan Goldin (University of California, Irvine). The clone for rNa_V1.6 (NM_019266.2) was prepared as described previously (Zhang et al. 2013a). Clones for rNa_vβ3 (NM_139097.3) and rNa_vβ4 (NM_001008880.1) were obtained from Lori Isom (University of Michigan, Ann Arbor). rNa_vβ1 and rNa_vβ2 DNA were linearized with NotI and transcribed with T7; rNa_vβ3 DNA was linearized with XbaI and transcribed with T7; and rNa_vβ4 DNA was linearized with BamHI and transcribed with T7.

Oocytes were harvested and prepared essentially as described previously (Cartier et al. 1996). Briefly, freshly excised oocytes were treated with 2.5 mg/ml collagenase A (Roche Diagnostics, Indianapolis, IN) in OR-2 (82.5 mM NaCl, 2.0 mM KCl, 1.0 mM MgCl₂, and 5.0 mM HEPES, pH 7.3) for 1–2 h on a rotary shaker at room temperature. Halfway through the collagenase treatment, the solution was exchanged with fresh collagenase solution. Following the enzyme treatment, oocytes were rinsed with OR-2 and incubated until used at 16°C in ND96 (96 mM NaCl, 2.0 mM KCl, 1.8 mM CaCl₂, 1.0 mM MgCl₂, and 5.0 mM HEPES, pH 7.3), supplemented with penicillin (100 U/ml) and streptomycin (0.1 mg/ml).

Injection of cRNA into oocytes was done as described previously (Wilson et al. 2011a; Zhang et al. 2013a). Briefly, a given oocyte was injected with 30–70 nl of 0.3 ng mNa_V1.6 or 117 ng rat cRNA in distilled water, without or with an equal weight of a rNa_Vβ cRNA. Oocytes were incubated at 16°C for 1- 6 days in ND96, supplemented with the aforementioned antibiotics.

Oocytes were two-electrode voltage clamped with an OC-725C amplifier (Warner Instruments, Hamden, CT), using 3 M KCl-filled microelectrodes (<0.5 MΩ resistance), essentially as described previously (Zhang et al. 2013a). A holding potential of −80 mV was used, and I_Na were induced every 20 s with a 50-ms depolarizing step to −10 mV. Current signals were filtered at 2 kHz, digitized at a sampling frequency of 10 kHz, and leak subtracted with a P/8 protocol using in-house software written in LabVIEW (National Instruments, Austin, TX). The recording chamber was a 4-mm diameter well (total volume of 30 μl), sunk in Sylgard (Dow Corning, Midland, MI), a silicone elastomer. Toxins were dissolved in ND96, and oocytes were exposed to toxin by applying 3 μl toxin solution (at 10× the final concentration) to a static bath with a pipettor and manually stirring the bath for a few seconds by gently aspirating and expelling a few microliters of bath fluid several times with the pipettor. A static bath was used to conserve toxin, and toxins were washed out by continuous perfusion with ND96, initially at a rate of 1.5 ml/min for 20 s and then at a steady rate of 0.5 ml/min.

Oocyte data were analyzed as follows. Percentage block of peak I_Na by toxin was determined by obtaining the average peak of greater than or equal to three control traces and the average peak of greater than or equal to three traces acquired at steady state in the presence of toxin and then dividing the latter by the former and multiplying by 100. Fitting of time-course data to a single exponential function was done with homemade software written with LabVIEW. The interaction of toxin with channel was assumed to be that of a simple bimolecular reaction whose kinetics are described by the equation, k_obs = k_on[toxin] + k_off, where [toxin] is toxin concentration, k_obs is the observed association rate constant, and k_off is the disassociation rate constant. The time course of peak I_Na was plotted before, during, and after exposure to toxin. The k_on was determined as follows: the onset of block at a given [toxin] was fit to a single exponential function to yield the k_obs, following which, k_on was obtained from the linear-regression slope of a k_obs vs. [toxin] plot for at least three different [toxins] (where each concentration was tested on greater than or equal to three oocytes), as described previously (Zhang et al. 2013a). The k_off was determined by fitting the toxin-washout curve to a single-exponential function; however, when recovery from block was very slow (<50% recovery after 20 min; i.e., k_off < 0.035/min), k_off was estimated from the level of recovery observed after 20 min of washing and assuming recovery followed a single exponential time course. Times longer than 20 min were not used to avoid error due to possible baseline drift. Each k_off value was the average of greater than or equal to nine oocytes.

Extracellular recording of CAPs from rat and mouse sciatic nerves.

Preparation of sciatic nerves and recordings from them were performed essentially as described previously (Fiedler et al. 2008; Wilson et al. 2011a). Briefly, sciatic nerves were dissected from adult male Sprague-Dawley rats and Swiss Webster or C57/BL6 mice, desheathed, and used within ∼2 h. A given nerve was placed in a multiwell Vaseline-gap chamber, made of Sylgard. Each well was 4 mm in diameter and ∼4 mm deep, with ∼1 mm wide partitions between adjacent wells. All wells contained mammalian Ringer's solution consisting of 140 mM NaCl, 5 mM KCl, 2 mM CaCl₂, 1.1 mM MgCl₂, and 5 mM HEPES, pH 7.4. The proximal end of the nerve was in well 1 and the distal end in either well 3 or 4. The portions of the nerve overlying the partitions were covered with Vaseline. The nerve was stimulated with wire electrodes in wells 1 and 2, and CAPs were recorded with a pair of wire electrodes in either wells 2 and 3 or 3 and 4, with the electrode in the lower-numbered well connected to the positive input of the preamplifier. A ground electrode was situated in well 2 or 3. All electrodes were platinum wires. Extracellular records were acquired with a differential capacitive-coupled preamplifier, band-pass filtered (1 Hz–1 kHz) and sampled at 4 kHz using in-house software written in LabVIEW. The amplitude of a 0.1- to 1-ms duration constant-voltage pulse was adjusted to provide a supramaximal stimulus that evoked both A- and C-CAPs, and the stimulus was applied once/minute. Normally, the nerve was exposed to toxin by replacing the solution in a middle well with toxin-containing Ringer's solution. In experiments with μO§-GVIIJ_SSC/G, the Ringer's solution also contained 0.1 mg/ml BSA to minimize nonspecific binding. [Separate experiments with oocytes showed that the presence of BSA did not affect the peptide's activity (not illustrated).] In the case of tests with δ-PVIA, which inhibits channel inactivation, the toxin was applied to the last well, i.e., the well with the recording electrode that fed into the negative input of the preamplifier, as described previously (Bulaj et al. 2001). Solutions in all wells were static and manually replaced every 10–15 min with fresh solutions. To minimize evaporation, the atmosphere immediately above the wells was exposed to a gentle stream of water-saturated air.

Following tests of μO§-GVIIJ_SSC on sciatic nerves, the tested (i.e., “used”) peptide was retrieved from the well and assayed for 1) functional activity on voltage-clamped oocytes expressing mNa_V1.6 (see above) and 2) structural integrity by HPLC (see below).

Whole-cell voltage-clamp recording of I_Na from acutely dissociated DRG neurons from rat.

DRG neurons of adult male Sprague-Dawley rats were dissociated and used as described previously (Zhang et al. 2013b). Briefly, ganglia were excised and treated with collagenase, followed by trypsin. Cells were mechanically dissociated by trituration, washed, and suspended in Leibovitz L-15 medium, supplemented with 14 mM glucose, 1 mM CaCl_2, 10% FBS, and penicillin/streptomycin. Dissociated neurons were kept in suspension at 4°C for up to 3 days (Blair and Bean 2002). Voltage-clamp recordings were performed with a MultiClamp 700A amplifier (Axon Instruments, Union City, CA) using a bath with a total volume of 100 μl, essentially as described previously (Zhang et al. 2013b). The extracellular solution contained 140 mM NaCl, 3 mM KCl, 1 mM MgCl₂, 1 mM CaCl₂, 0.1 mM CdCl₂, and 20 mM HEPES, pH 7.3. Patch pipettes had resistances of <2 MΩ and contained 140 mM CsF, 10 mM NaCl, 1 mM EGTA, and 10 HEPES, pH 7.3; series resistance compensation was >80%. After the achievement of whole-cell clamp conditions, recordings were not initiated until the holding current had settled, which required >10 min. The contribution of Na_V1.9, relative to that of Na_V1.8, to the TTX-resistant current of DRG neurons is minimized by such a settling period (Choi et al. 2006). The membrane potential was held at −80 mV, and I_Na were elicited with a 50-ms step to 0 mV, applied every 20 s. Current signals were low-pass filtered at 3 kHz, digitized at a sampling frequency of 10 kHz, and leak subtracted by a P/6 protocol using in-house software written in LabVIEW.

Toxins were dissolved in extracellular solution and applied to the clamped neuron by simple bath exchange that involved manually applying, with a pipette, toxin solution (150 μl) at one end of the boat-shaped, 100-μl chamber, while simultaneously withdrawing solution at the other end of the chamber over a time span of <20 s. The patch electrode was used to lift the cell from the chamber bottom and position the cell near the upstream half of the chamber to ensure that the cell was fully exposed to the introduced toxin solution. Toxin exposures were conducted in a static bath to conserve toxin. The level of TTX-resistant I_Na of each DRG cell was determined by perfusion with 1 μM TTX following tests with μO§-GVIIJ_SSG, which does not block rNa_V1.8 expressed in oocytes (Gajewiak et al. 2014).

Use of animals in this study followed protocols approved by the University of Utah's Institutional Animal Care and Use Committee that conform to the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

All electrophysiological experiments were conducted at room temperature (∼22°C).

HPLC of used μO§-GVIIJ_SSC.

Qualitative and quantitative analyses of used μO§-GVIIJ_SSC (see above) were performed with an analytical C₁₈ Vydac reversed-phase HPLC column (218TP54, 250 mm × 4.6 mm, 5 μm particle size) at 23°C. Solvents consisted of 0.1% (v/v) trifluoroacetic acid in either water (solvent A) or 90% aqueous acetonitrile (solvent B). The sample was eluted with solvent A and a linear gradient of 15–45% solvent B. The optical absorbance of the eluent was monitored at 220 nm.

RESULTS

mNa_V1.6, expressed in oocytes, is blocked by μ-contoxins BuIIIB and SxIIIA more potently than rNa_V1.6.

In previous experiments, we tested μ-BuIIIB and μ-SxIIIA against mNa_V1.6 expressed in X. laevis oocytes without any Na_Vβ-subunit coexpression, and they had similar IC₅₀ (1.8 and 0.57 μM, respectively) (Wilson et al. 2011a). All of the experiments described below that involved coexpression with Na_Vβ-subunits used β-subunits from rat, because clones from mouse were unavailable to us. Thus we tested μ-BuIIIB and μ-SXIIIA against rNa_V1.6 and mNa_V1.6, coexpressed with rNa_Vβ1, and observed that although both peptides had similar affinities for a given channel, each had a higher affinity for mNa_V1.6 than rNa_V1.6 (Fig. 1), where the higher affinity of μ-BuIIIB for mNa_V1.6 is a consequence of a larger k_on and smaller k_off (Table 1). Values of k_on for μ-SXIIIA were too large to be measured with our oocyte system; however, the k_off of μ-SxIIIA for the rat channel is similar to that for the mouse channel (Table 1), indicating that the higher affinity of μ-SxIIIA for mNa_V1.6 results from a larger k_on for the mouse, over rat, channel.

Fig. 1.

Block by μ-BuIIIB and μ-SxIIIA of mouse and rat α-subunit of voltage-gated sodium channels (Na_V1.6; mNa_V1.6 and rNa_V1.6, respectively), coexpressed with rat β-subunit (rNa_Vβ1) in Xenopus laevis. Oocytes were prepared and voltage clamped as described in materials and methods. The holding potential was −80 mV, and voltage-gated sodium currents (I_Na) were induced by a 50-ms step to −10 mV, applied every 20 s. A–D: each pair of panels shows example time course of block (top; where bars represent when toxin was present) and sample responses (bottom; where control traces are gray and in-toxin traces are black). Block of mNa_V1.6 by 10 μM μ-BuIIIB (A) or μ-SxIIIA (B) and rNa_V1.6 by 10 μM μ-BuIIIB (C) or μ-SxIIIA (D). E: percentage block as a function of toxin concentration ([Toxin]) of mNa_V1.6 (left pair of curves) and rNa_V1.6 (right pair of curves) with μ-BuIIIB (closed symbols) and μ-SxIIIA (open symbols). Data points are mean ± SD (n ≥ 4 oocytes). Continuous, solid (μ-SxIIIA) and dashed (μ-BuIIIB) curves are fits to the Langmuir adsorption isotherm; IC₅₀ values derived from these fits are presented in Table 1.

Table 1.

Comparison of block by μ-SxIIIA and μ-BuIIIB of rat vs. mouse Na_V1.6, expressed with rat Na_Vβ1 in Xenopus laevis oocytes^*

	μ-SxIIIA			μ-BuIIIB
	kon, μM−1min−1	koff, min−1	IC50, μM	kon, μM−1min−1	koff, min−1	IC50, μM
Channel
rNaV1.6 + rβ1	NA†	1.55 ± 0.91	19.89 ± 3.24‡	0.035 ± 0.002	0.80 ± 0.28	30.36 ± 10.80‡
mNaV1.6 + rβ1	NA†	1.10 ± 0.54	0.48 ± 0.06‡	0.156 ± 0.024	0.17 ± 0.09	0.32 ± 0.07‡

^*Values are mean ± SD [n ≥ 9 oocytes for association rate constant (k_on) and IC₅₀ values; n ≥ 3 oocytes for disassociation rate constant (k_off) values], determined as described in materials and methods. ^†NA, values not available because observed time courses of block at the toxin concentrations tested were too fast to quantify accurately.

^‡IC₅₀ values from Fig. 1E. rNa_V1.6, rat α-subunit of voltage-gated sodium channels; Na_Vβ1, β-subunit of voltage-gated sodium channels; rβ1, rat β1; mNa_V1.6, mouse Na_V1.6.

The block by μ-SxIIIA of mNa_V1.6, coexpressed with various combinations of the four Na_Vβ-isoforms, was examined (Table 2). Coexpression with Na_Vβ1 or -β3 minimally affected affinity. In contrast, coexpression with Na_Vβ2 or -β4 increased the IC₅₀ >10-fold. This increase in IC₅₀ observed with unary coexpression of Na_Vβ2 or -β4 was largely “reversed” with binary coexpression of Na_Vβ2 or -β4 with Na_Vβ1 or -β3 (Table 2).

Table 2.

Block by μ-SxIIIA of mNa_V1.6 expressed alone or coexpressed with various combinations of the 4 isoforms of rNa_Vβ-subunits

mNaV1.6	koff, min−1*	IC50, μM*	% Block by 10 (or 30) μM μ-SxIIIA†
Alone	0.18 ± 0.06‡	0.57 ± 0.08‡	95 (98)
+β1	1.10 ± 0.14§	0.48 ± 0.04§	95 (98)
+β2	4.92 ± 0.72	7.74 ± 0.12	56 (80)
+β3	4.80 ± 1.30	0.44 ± 0.01	96 (99)
+β4	3.80 ± 1.21	11.21 ± 0.30	47 (73)
+β1+β2	3.53 ± 1.40	2.32 ± 0.03	82 (93)
+β1+β4	2.47 ± 0.35	1.05 ± 0.05	91 (97)
+β3+β2	4.21 ± 1.59	1.15 ± 0.01	90 (96)
+β3+β4	4.74 ± 0.80	1.08 ± 0.02	90 (97)

^*Values are mean ± SD (n ≥ 9 oocytes for IC₅₀ values; n ≥ 3 oocytes for k_off values). Values of k_on are not available because the observed association rate constant at the concentrations tested was too large to measure.

^†Predicted percentage of channels blocked from the Langmuir equation, % block = 100%/(1 + C/IC₅₀), where C = 10 μM (or 30 μM for values in parentheses).

^‡From Wilson et al. (2011a).

^§From Table 1.

μO§-GVIIJ_SSC and -GVIIJ_SSG block mNa_V1.6, expressed in oocytes, except when Na_Vβ2 or -β4 is coexpressed.

Two synthetic derivatives of the recently discovered μO§-GVIIJ were tested, GVIIJ_SSC and GVIIJ_SSG, where the subscripts indicate that cysteine or glutathione, respectively, was disulfide bonded to Cys24 of the peptide (see materials and methods). mNa_V1.6 expressed alone was readily blocked by μO§-GVIIJ_SSC or μO§-GVIIJ_SSG (Fig. 2, A and J). μO§-GVIIJ_SSC also readily blocked mNa_V1.6, coexpressed with either Na_Vβ1 or -β3 (Fig. 3, B and D); in contrast, coexpression with Na_Vβ2 or -β4, either alone or in binary combination with Na_Vβ1 or -β3, protected the channels against block by the peptide (Fig. 2, E–I). The kinetic constants of the block by μO§-GVIIJ_SSC are tabulated in Table 3.

Fig. 2.

μO§-GVIIJ_SSC (where subscript refers to cysteinylated) readily blocks mNa_V1.6, expressed alone or coexpressed with rNa_Vβ1 or -β3; however, coexpression with rNa_Vβ2 or -β4 protects mNa_V1.6 against block by μO§-GVIIJ_SSC. Oocytes were voltage clamped, as described in Fig. 1. A–J: there are 10 pairs of panels. Top: representative time course of block before, during, and after exposure to GVIIJ_SSC (A–I) or GVIIJ_SSG (where subscript refers to glutathionylated; J), where the bar represents when peptide was present (thin black bars, 1 μM; thick black bars, 33 μM). Bottom: sample responses before (light traces) and during (dark traces) toxin application. First and last pair of panels represent mNa_V1.6 expressed alone (A and J), and the remaining 8 pairs of panels are labeled with the rNa_Vβ-subunits that were coexpressed with mNa_V1.6 (B–I). GVIIJ_SSC (1 μM) readily blocked mNa_V1.6 expressed alone (A) or coexpressed with either Na_Vβ1 (B) or Na_Vβ3 (D); in contrast, 33 μM GVIIJ failed to block whenever Na_Vβ2 or -β4 was coexpressed in either unary fashion (C and E) or binary combination with Na_Vβ1 (F and G) or -β3 (H and I). (Kinetic constants derived from replicate experiments are presented in Table 3). J: block of mNa_V1.6 (no Na_Vβ-subunit coexpressed) by 10 μM GVIIJ_SSG; time course (top; bar represents when peptide was present) and representative traces (bottom).

Fig. 3.

Susceptibilities of large, myelinated A-fiber compound action potentials (A-CAPs) and small, unmyelinated C-fiber CAPs (C-CAPs) of the mouse sciatic nerve to each of 7 μ-conotoxins (10 μM) or TTX (0.1 μM). CAPs were evoked by electrical stimulation and recorded as described in materials and methods. A–H: top 2 plots in each 4-plot panel show time course of normalized peak-to-peak amplitudes of A-CAPs (left) and C-CAPs (right), simultaneously recorded from the sciatic nerve, where the bars above each plot indicate when toxin was present; the x-axis scale (representing time in minutes), shown in left plots, applies to all top plots. Data points represent mean ± SE values (n = 3 sciatic nerves). Discontinuities in plots occurred when solutions in wells were refreshed (see materials and methods). Bottom 2 plots in each 4-plot panel show representative traces of A-CAPs (left) and C-CAPs (right). Traces obtained in control solution are gray, and in those, the presence of 10 μM μ-conopeptide (A–G) or 0.1 μM TTX (H) is black and either attenuated or largely unchanged compared with controls. A- and C-CAPs were captured in the same sweep but for clarity, are displayed with different time bases. Note stimulus artifact at the start of each A-CAP trace. The difference in latencies (i.e., time delays between stimulus and response) of A- and C-CAPs varied depending on length of nerve, and for illustration purposes, a fixed-space gap was placed between the end of a given A-CAP trace and the beginning of the following C-CAP trace. A quantitative summary of these results is presented in Table 4. I: time course of block of mouse A-CAPs by 3 different concentrations of μ-SxIIIA, showing that a concentration of 30 μM μ-SxIIIA blocks 95% of mouse A-CAPs (data for 10 μM peptide were obtained from F).

Table 3.

Block by μO§-GVIIJ_SSC of mNa_V1.6 expressed alone or coexpressed with various combinations of the 4 isoforms of rNa_Vβ-subunits

mNaV1.6	kon, μM−1/min−1*	koff, min−1*	Kd, μM†
Alone	1.15 ± 0.08	0.069 ± 0.026	0.060 ± 0.023
+β1	1.42 ± 0.06	0.042 ± 0.023	0.029 ± 0.016
+β2	NA‡	NA‡	NA‡
+β3	1.69 ± 0.08	0.093 ± 0.025	0.055 ± 0.015
+β4	NA‡	NA‡	NA‡
+ β1+β2	NA‡	NA‡	NA‡
+ β1+β4	NA‡	NA‡	NA‡
+ β3+β2	NA‡	NA‡	NA‡
+ β3+β4	NA‡	NA‡	NA‡

^*Values are mean ± SD (n ≥ 9 oocytes for k_on values; n ≥ 3 oocytes for k_off values).

^†K_d, dissociation constant; determined from k_off/k_on.

^‡NA values not available because no block was observed in n ≥ 3 oocytes. Representative responses are illustrated in Fig. 2. SSC, cysteinylated analog.

Block by μ-conotoxins of A- and C-CAPs in the mouse sciatic nerve.

We previously observed that neither A- nor C-CAPs in the rat sciatic nerve are blocked by μ-SxIIIA or μ-BuIIIB (Wilson et al. 2011a). However, our results above suggest that if Na_V1.6 were a major Na_V1-isoform of A-fibers in mouse sciatic nerve, then both peptides should be able to block mouse A-CAPs. This was indeed observed (Fig. 3, F and G; Table 4).

Table 4.

Kinetics and levels of block of A- and C-CAPs of mouse sciatic nerve by 7 μ-conopeptides; comparison with expected block of mNa_V1.6 expressed in X. laevis oocytes

	A-CAPs		C-CAPs		mNaV1.6
Toxin (μM)	% Block*	t1/3, min†	% Block*	t1/3, min†	% Block‡
μ-SmIIIA (10)	100 ± 0§	2.2	62.6 ± 2.2§	1.5	98
μ-KIIIA (10)	100 ± 0§	2.8	38.7 ± 4.3§	12	98
μ-SIIIA (10)	100 ± 0§	15	41.6 ± 2.1§	37	93
μ-GIIIA (10)	100 ± 0§	1.1	0.8 ± 2.2	N/A	94
μ-PIIIA (10)	100 ± 0§	<1	2.6 ± 4.7	N/A	99
μ-SxIIIA (10)	69 ± 3.9§	<1	0.8 ± 4.24	N/A	95
μ-BuIIIB (10)	94.1 ± 4.0§	6	17.5 ± 8.0¶	N/A	85
TTX (0.1)	100 ± 0§	<1	82.4 ± 4.7§	1

^*Block percentage value (mean ± SE; n = 3), determined from average of last 4 traces in the presence of peptide compared with average of 5 traces just before exposure to peptide.

^†t_1/3, time to block by 33.3%. Significantly different from no block;

^¶P < 0.05,

^§P < 0.001 (1-tail Student's t-test, assuming % block cannot be <0).

^‡Expected block percentage of mNa_V1.6 with no Na_Vβ-subunit coexpressed in X. laevis oocytes from Supplemental Table 5 of Wilson et al. (2011a). A-CAPs, large, myelinated A-fiber compound action potentials; C-CAPs, small, unmyelinated C-fiber CAPs.

A-CAPs in the mouse sciatic nerve were also blocked by five other μ-contoxins (Fig. 3, A–E, and Table 4) that were previously shown to block mNa_V1.6 expressed in oocytes (Wilson et al. 2011a). Two of these, μ-SmIIIA and μ-KIIIA, not only blocked mouse sciatic nerve A-CAPs but also rapidly blocked C-CAPs (Fig. 3, A and B), just as they do A- and C-CAPs of the rat sciatic nerve (Wilson et al. 2011a). μ-SIIIA slowly blocked mouse C-CAPs (Fig. 3C), whereas rat C-CAPs were resistant to μ-SIIIA (Wilson et al. 2011a). Thus it is possible that mNa_V1.7 may be more susceptible than rNa_V1.7 to μ-SIIIA. Although μ-BuIIIB may block C-CAPs, it does so only poorly if at all (Fig. 3G and Table 4).

At 10 μM, all of the μ-conotoxins in Fig. 3 (and also listed in Table 4) blocked mouse A-CAPs essentially completely, except μ-SxIIIA (Fig. 1C). To examine this further, a threefold-higher concentration of μ-SxIIIA (30 μM) was tested; it blocked the A-CAP at least 95% (Fig. 3I), consistent with a pharmacologically homogeneous population of channels underlying A-CAPs.

TTX (0.1 μM) blocked 100% of the A-CAPs and ∼80% of the C-CAPs (Table 4), where the conduction velocity of the highly attenuated C-CAPs was decreased markedly (Fig. 3H). The residual amplitude of C-CAPs that persists presumably reflects the presence of Na_V1.8 and/or -1.9, which are TTX resistant, in C-fibers. These TTX results essentially mirror those observed with A- and C-CAPs of the rat sciatic nerve (Wilson et al. 2011a).

A- and C-CAPs in mouse and rat sciatic nerves are not blocked by μO§-GVIIJ_SSC or -GVIIJ_SSG.

Neither μO§-GVIIJ_SSC nor μO§-GVIIJ_SSG was able to block A- and C-CAPs in mouse and rat sciatic nerves at 33 μM (Fig. 4, A and B), a concentration well beyond that necessary to block the majority of the I_Na of mNa_V1.6, provided it wasn't coexpressed with Na_Vβ2 or -β4 (Fig. 2 and Table 3). In addition, inspection of Fig. 4 also shows that μO§-GVIIJ_SSC/G produced no increase in the latency of A- or C-CAP, indicating no decrease in the conduction velocity of action potentials in A- or C-fibers.

Fig. 4.

A- and C-CAPs of rat and mouse sciatic nerves were not blocked by 33 μM μO§-GVIIJ_SSC or -GVIIJ_SSG. CAPs were evoked, recorded, and illustrated, essentially as in Fig. 3. A and B: top 2 plots in each 4-plot panel show time course of normalized peak-to-peak amplitudes of simultaneously recorded A-CAPs (left) and C-CAPs (right) from rat (A) and mouse (B) sciatic nerves, where the bars above each plot indicate when 33 μM μO§-GVIIJ_SSC was present (note absence of data point near minute 5, when μO§-GVIIJ_SSC was being applied), and arrows depict when 10 μM μ-SmIIIA was introduced (note compressed time base for data points representing responses in μ-SmIIIA). Data points for μO§-GVIIJ_SSC represent mean ± SE values (n = 3 sciatic nerves), and those for μ-SmIIIA are from single nerves (i.e., n = 1). Bottom 2 plots in each 4-plot panel show example traces of A- and C-CAPs obtained in control solution (dashed traces), in 33 μM μO§-GVIIJ_SSC (solid traces, essentially same as control traces), and in 10 μM SmIIIA (dotted traces, where A-CAP was totally blocked, while C-CAP was dramatically slowed and attenuated). The shaded area at the start of the mouse A-CAP trace contains the stimulus artifact and can be ignored. Toward the end of a rat nerve trial, 33 μM GVIIJ_SSG was also applied, and still, no block of A- and C-CAPs was seen (not illustrated). C: the “used” GVIIJ_SSC solution from a rat trial was retrieved at the end of the experiment and subjected to analysis by HPLC (see materials and methods). The elution profile consisted of a single peak with the same retention time as unused GVIIJ_SSC (Gajewiak et al. 2014), showing that the peptide remained intact, and integration under the peak indicated that ∼0.8 nmol was applied to the column, a value 29% lower than the expected 1.1 nmol. A₂₂₀, absorbance at 220 nm.

To examine whether the peptide was degraded or absorbed by the tissue to which it was exposed, the used toxin solution was retrieved and examined for structural integrity by subjecting it to HPLC analysis (Fig. 4C), as well as for functional activity against mNa_V1.6 expressed in oocytes, which was performed as follows. A sample of used GVIIJ_SSC solution was diluted 10-fold with ND96 and assayed on an oocyte expressing mNa_V1.6. The k_obs of I_Na was 3.0 min⁻¹, which (after accounting for the 10-fold dilution) is that expected of a GVIIJ_SSC concentration of 26 μM, a value 21% lower than the starting 33 μM. The slightly (20–30%) lower-than-expected values obtained with the HPLC and oocyte assays can be explained by dilution due to siphoning of Ringer's solution from adjacent wells during retrieval of the contents from the toxin-containing well. Thus by both structural and functional assays, the used toxin proved to be intact and largely recoverable.

δ-PVIA affects A-CAPs of the mammalian sciatic nerve, like members of all other families of conotoxins that target VGSCs except μO§-conotoxins.

As mentioned previously, there are five families of conotoxins that target VGSCs. Until now, all families, except μO§- and δ-conotoxins, have been shown to affect A-CAPs of the mammalian sciatic nerve (see discussion). Earlier experiments with δ-PVIA and -SVIE showed that these peptides prolonged the action potentials in peripheral nerves of frogs (Rana pipiens) (Bulaj et al. 2001; West et al. 2005), and here, we examined whether δ-PVIA could be shown also to act on the mammalian (specifically, rat) sciatic nerve. Indeed, 10 μM δ-PVIA greatly prolonged the A-CAP (Fig. 5). The prolonged A-CAPs persisted longer than the latency of the C-CAPs and therefore, rendered accurate characterization of the latter problematical, so the effects of δ-PVIA on C-CAPs remain to be determined. The effect of δ-PVIA on A-CAPs was very robust and distinctive and closely resembled that seen with δ-SVIE on A-CAPs of frog nerve [see Fig. 3B in Bulaj et al. (2001)]. The profound potentiation and prolongation of the upward phase of A-CAPs induced by δ-PVIA, illustrated in Fig. 5, were observed in duplicate (two out of two) trials but never in traces of countless trials without the peptide [e.g., see traces in Fig. 4, as well as Fig. 3, of Wilson et al. (2011a)]. This result supports the notion that peptide inaccessibility is an unlikely cause for the lack of activity of μO§-GVIIJ_SSC/G on the sciatic nerve, as recounted in discussion.

Fig. 5.

δ-Conotoxin PVIA (δ-PVIA) greatly enhances and prolongs the A-CAP of the rat sciatic nerve. Recording protocol differed slightly from that used in Figs. 3 and 4, in that the toxin was added to the downstream recording well, which contained the electrode that fed into the negative input of the differential recoding preamp, as described in materials and methods. A: A-CAP, followed by much smaller C-CAP (note “bump” at ∼80 ms), before (dashed traces) and after (solid traces) exposure to 10 μM δ-PVIA. B: same as A but showing A-CAP on an expanded time axis; note that only the later (upward-going) phase of the biphasic A-CAP was potentiated by δ-PVIA, consonant with the downstream location (relative to the direction of action potential propagation) of the well in which the toxin was placed. The 1-ms stimulus was applied at 20 ms.

μO§-GVIIJ_SSG blocks I_Na of small, but not large, dissociated rat DRG neurons.

The inability of μO§-GVIIJ_SSC/G to block CAPs raises the question of whether these peptides had any activity on endogenous VGSCs. To examine this issue, three small and three large dissociated rat DRG neurons were whole-cell patch clamped as before (Zhang et al. 2013b) and subjected to 10 μM μO§-GVIIJ_SSG. The peptide blocked, albeit partially, TTX-sensitive I_Na of all three small neurons; in contrast, all three large neurons were insensitive to the peptide (Fig. 6). Similar tests with μO§-GVIIJ_SSC remain to be performed.

Fig. 6.

I_Na of small, but not large, dorsal root ganglia neurons are blocked by 10 μM μO§-GVIIJ_SSG. Acutely dissociated neurons were whole-cell patch clamped, as described in materials and methods. The holding potential was −80 mV, and I_Na was induced by a 50-ms step to 0 mV, applied every 20 s. A: percentage block by GVIIJ_SSG of TTX-sensitive I_Na of 3 small and 3 large neurons; neuron sizes were determined quantitatively by membrane capacitance (see x-axis) and qualitatively by visual inspection. B: 4-plot panel showing representative traces of small (top left) and large (top right) neurons before (light traces) and during (dark traces) exposure to GVIIJ_SSG. Peak I_Na of these neurons plotted as a function of time (bottom); bars represent when neuron was exposed to GVIIJ_SSG. Illustrated are results of the 12.6-pF small neuron (bottom left) and 45.8-pF large neuron (bottom right). The I_Na of all neurons was blocked essentially completely by 0.1 μM TTX, except for the 11.1-pF neuron, which was blocked by 60% (not illustrated).

DISCUSSION

In this report, we examined the effects primarily of two conotoxin families, μ- and μO§-conotoxins, on VGSCs in three preparations: 1) X. laevis oocytes exogenously expressing Na_V1.6, with or without various Na_Vβ-isoforms; 2) A- and C-fibers of rodent sciatic nerves; and 3) soma of dissociated small and large neurons of rat DRG.

μ-Conotoxin sensitivities of rNa_V1.6 vs. mNa_V1.6.

μ-BuIIIB and μ-SxIIIA blocked Na_V1.6 of rat and mouse with IC₅₀ values that differ by >50-fold (Fig. 1E and Table 1). This is in contrast to two other μ-conotoxins for which affinity data for both rat and mouse are available; namely, μ-SmIIIA and μ-PIIIA, each of whose dissociation constants for rNa_V1.6 vs. mNa_V1.6 differ by a factor of only 2.3 or less (Wilson et al. 2011a; Zhang et al. 2013a). rNa_V1.6 and mNa_V1.6 differ in 10 residues, six of which are in the extracellular portions of the channel. Of these, three are in the pore loop regions: one in DII and two in DIII. The residue in question in DII is located near the N-terminal end of the S5–S6 (pore) loop: Asn907 in rNa_V1.6 and Ser907 in mNa_V1.6. The homologous residue in rNa_V1.4 is Ala728, and the activity of μ-conopeptide GIIIA, a potent blocker of rNa_V1.4 (Cruz et al. 1985), has been tested extensively against mutants of rNa_V1.4 by several investigators, and a subset of those studies shows that mutations of Ala728 do influence the binding of μ-GIIIA and its congener, μ-GIIIB (Chahine et al. 1998; Li et al. 2003). Furthermore, it was recently shown that the block of rNa_V1.4 by μ-SIIIA is reduced significantly by an A728N mutation (Leipold et al. 2011). By extension, the single-residue difference in DII between rNa_V1.6 and mNa_V1.6 may be largely responsible for their different sensitivities to μ-BuIIIB and μ-SxIIIA.

Modulation of the sensitivity of mNa_V1.6 to μ-SxIIIA by coexpression of rNa_Vβ-subunits.

Investigations of the pharmacological consequences of coexpression of Na_Vβ-subunits have been limited, but several previous studies demonstrate that coexpression of Na_Vβ-subunits can affect the affinity of agents targeted at the Na_V1- or α-subunit (Gajewiak et al. 2014; Tan et al. 2011; Wilson et al. 2011b; Zhang et al. 2013a). Thus it is no surprise that the block of mNa_V1.6 by μ-SxIIIA was also modulated by coexpression of Na_Vβ-subunits. mNa_V1.6, coexpressed without or with various rat β-subunits, had the following overall sequence of IC₅₀ values: no beta ≈ +β1 ≈ +β3 ≈ 0.5 μM < +β1+β4 ≈ +β3+β4 ≈ +β3+β2 ≈ 1 μM < +β1+β2 ≈ 2 μM < +β2 ≈ 8 μM < +β4 ≈ 11 μM (Table 2). This sequence of affinities, produced by unary and binary coexpression of the various Na_Vβ-subunits, is also observed in the block of rNa_V1.7 by μ-SmIIIA (Zhang et al. 2013a); that is, unary coexpression of Na_Vβ1 or -β3 has minimal effects, and that of Na_Vβ2 or -β4 has the largest effects, whereas binary coexpression of Na_Vβ1 or -β3 with Na_Vβ2 or -β4 tempers the effects observed with unary coexpression of Na_Vβ2 or -β4.

Despite the congruity of results described immediately above, it should be noted that there are differences in the sequences of a given rNa_Vβ-isoform and that from mNa_V1.6—there is a two-residue difference for β1, one for β3, six for β2, and five for β4—so the results from coexpression of an Na_V1-isoform from one rodent species with an Na_Vβ-isoform from a another species should be viewed with caution.

Modulation of the sensitivity of mNa_V1.6 to μO§-GVIIJ_SSC by coexpression of rNa_Vβ-subunits.

The recently discovered μO§-GVIIJ is the charter member of a new, fifth family of conotoxins that target VGSCs. Until now, it has been tested only on rat VGSCs expressed in oocytes, where it readily blocks rNa_V1.1, -1.2, -1.3, -1.4, -1.6, and -1.7 (Gajewiak et al. 2014), all of which happen to be sensitive to TTX. It should be noted that the TTX sensitivity is largely dictated by an aromatic residue in site 1 near the ion-selectivity filter (Santarelli et al. 2007), whereas the μO§-GVIIJ sensitivity is dictated by a Cys residue in the newly described site 8 between S5 and SS5 (the latter is the proximal limb of the pore loop) of DII (Gajewiak et al. 2014).

By and large, μO§-GVIIJ_SSC and μO§-GVIIJ_SSG behave similarly, both in previous (Gajewiak et al. 2014) and present experiments. A subtle difference is that the on rate of μO§-GVIIJ_SSC is apparently larger than that of μO§-GVIIJ_SSG, which is evident in Fig. 2 (compare pairs in Fig. 2A with Fig. 2J, keeping in mind the difference in peptide concentrations). This is also the case with rNa_V1.2 and -1.7, where the k_on of μO§-GVIIJ_SSC is approximately threefold larger than that of μO§-GVIIJ_SSG (Gajewiak et al. 2014). Cysteine is smaller than glutathione, and this may explain the difference in k_on. A detailed examination of this issue is under way with derivatives of μO§-GVIIJ, where the moiety disulfide bonded to Cys24 of the peptide varies over a range of sizes and charges (unpublished observations).

A stellar feature of μO§-GVIIJ_SSC/G is that coexpression with Na_Vβ2 or -β4 renders otherwise-susceptible rNa_V1s resistant to the peptide (Gajewiak et al. 2014). We now show that this modulation by coexpression of Na_Vβ2 or -β4 also applies to mNa_V1.6 (Fig. 2 and Table 3). Furthermore, the protection of mNa_V1.6 against μO§-GVIIJ_SSC/G block by coexpression with Na_Vβ2 or -β4 persists in the face of binary coexpression with Na_Vβ1 or -β3; i.e., the effects of Na_Vβ2 or -β4 coexpression dominate those of Na_Vβ1 or -β3 (Table 3). Similarly, dominance of the coexpression of Na_Vβ2 and -β4 over that of Na_Vβ1 was observed with rNa_V1.2 (Gajewiak et al. 2014). This is unlike the effects of Na_Vβ2 or -β4 coexpression on the block of mNa_V1.6 by μ-SxIIIA, which are attenuated by binary coexpression with Na_Vβ1 or -β3 (Table 2). As mentioned in introduction, Na_Vβ2 and -β4 are disulfide bonded to the α-subunit. The binding site of μO§-GVIIJ on Na_V1, site 8, has a Cys residue with which we hypothesized the peptide can form a disulfide bond; furthermore, we speculated that Na_Vβ2 and -β4 may protect the channel against block by being disulfide bonded with the Cys at site 8 (Gajewiak et al. 2014). Further investigation of this possibility is underway.

It should be emphasized that the modulatory effect of coexpression of α- with β-subunits does not necessarily mean that the modulation results from the subunits' physical association per se; for example, Na_Vβ1 coexpression can affect the glycosylation of the α-subunit (Laedermann et al. 2013). Thus in principle, the alteration in the pharmacology of a VGSC could result from the coexpression of the β-subunit altering the processing of the channel.

Effects of μ-BuIIIB, μ-SxIIIA, and five other μ-conotoxins on A- and C-CAPs of mouse sciatic nerve are consistent with Na_V1.6 and -1.7 mediating action potentials in A- and C-fibers, respectively.

In previous experiments, we used a panel of μ-conotoxins and concluded that Na_V1.6 and -1.7 were the major Na_V1-isoforms responsible for the conduction of action potentials in A- and C-fibers, respectively, of rat sciatic nerve (Wilson et al. 2011a). Results in Fig. 3 and Table 4 lead us to conclude the same for the mouse sciatic nerve. We do not have a clone of mNa_V1.7, so we assume that mNa_V1.7 behaves similar to rNa_V1.7 toward the tested μ-conotoxins, with the possible exception of μ-SIIIA, which we speculated in results may block rNa_V1.7 better than rNa_V1.7. This issue can be resolved when an mNa_V1.7 clone becomes available.

The safety factor for action potential conduction can be defined as the current generated during an action potential divided by the threshold current necessary for the action potential to be propagated (Fern and Harrison 1993) or for our purposes, the density of sodium channels available for activation during an action potential divided by the minimum density of channels necessary for action potential conduction. In view of the safety factor, normally, a disproportionate fraction of sodium channels must be blocked before the action potential is blocked. For example, the IC₅₀ values of μ-SxIIIA in blocking rNa_V1.6+β1 and mNa_V1.6+β1 are 20 and 0.5 μM, respectively (Table 1). Given these IC₅₀ values, the Langmuir equation (see Table 2) predicts that 10 μM SxIIIA would block 95% and 33% of the mouse and rat channels, respectively. On the other hand, our experiments show that 10 μM μ-SxIIIA blocked ∼70% of the A-CAPs in mouse (Table 4) and essentially 0% of those in rat [see Table 3 of Wilson et al. (2011a)]. We do not know the safety factor for the conduction of A-CAPs in our experiments; however, the predicted block of Na_V1.6-containing VGSCs vs. the observed block of A-CAPs by 10 μM SxIIIA can be qualitatively reconciled if we assume that the safety factor is approximately two for both mouse and rat nerves; that is, >50% of the channels must be blocked for A-CAPs to fail. The disparity in the block of channels vs. action potentials would be larger, the larger the safety factor. This analysis requires several assumptions and does not rule out the possible contribution of minor Na_V1-isoforms to A-CAPs.

Inability of μO§-GVIIJ_SSC/G to block A- and C-CAPs in rat and mouse sciatic nerves suggests that channels expressed in A- and C-fibers are associated with Na_Vβ2 or -β4.

GVIIJ_SSC and GVIIJ_SSG blocked neither A- nor C-CAPs in rat and mouse sciatic nerves (Fig. 4). In principle, there are four relatively simple alternative reasons why no block was observed with μO§-GVIIJ_SSC/G; namely, the peptide is 1) physically unable to reach to the axon surface where VGSCs reside; 2) absorbed by the tissue preparation; 3) rapidly degraded; or 4) unable to block endogenously expressed VGSCs. To address the first possibility, at the end of a trial with each of a mouse and rat sciatic nerve, the nerve was exposed to 10 μM μ-SmIIIA, which readily blocked both A- and C-CAPs (Fig. 4), as was invariably observed in multiple experiments with sciatic nerves of both mouse (Fig. 3A and Table 4) and rat (Wilson et al. 2011a). In addition to μ-SmIIIA, many other μ-conotoxins are able to block A- and/or C-CAPs in sciatic nerves of mouse (Fig. 3 and Table 4) and rat (Wilson et al. 2011a). Furthermore, aside from μO§-GVIIJ_SSC/G, action potentials in A- and/or C-fibers of the mammalian sciatic nerve are susceptible to members of all other families of conotoxins that target VGSCs, i.e., μO-conopeptides (Bulaj et al. 2006), ι-conopeptides (Fiedler et al. 2008), and δ-conopeptides (Fig. 5). It should be noted that μO- and δ-conopeptides are more hydrophobic than μO§-GVIIJ, whereas ι-conotoxins are larger than μO§-GVIIJ. Thus our overall results suggest that a peptide-accessibility issue is unlikely with μO§-GVIIJ_SSC/G.

To address the second and third possibilities, after one of the trials in Fig. 3, the used μO§-GVIIJ_SSC was recovered and observed to be structurally intact (Fig. 4C) and functionally active (see results regarding Fig. 4), which shows that μO§-GVIIJ_SSC was not appreciably absorbed or degraded during exposure to tissue.

To address the fourth possibility, we examined dissociated rat DRG neurons. μO§-GVIIJ_SSG partially blocked the I_Na of small neurons, whereas the I_Na of large neurons were spared (Fig. 6). This showed that at least some endogenous channels are susceptible to the peptide. A- and C-CAPs arise from fast- and slow-conducting axons of neurons with, respectively, large and small cell somas in DRG (Harper and Lawson 1985); thus the resistance of the I_Na of large cells to μO§-GVIIJ_SSG (Fig. 6) is consistent with the resistance of A-CAPs to the peptide (Fig. 4). A-Fibers are myelinated, so one would predict that VGSCs at nodes of Ranvier are associated with Na_Vβ2 or -β4. Furthermore, since DRG neurons with small soma diameters give rise to unmyelinated C-fibers responsible for C-CAPs, the resistance of C-CAPs to μO§-GVIIJ_SSC/G suggests that in contrast to VGSCs expressed on the soma of small neurons, VGSCs in the bulk of its axons are associated with Na_Vβ2 or -β4.

The Na_Vβ4-subunit is involved in resurgent currents (Bant and Raman 2010; Lewis and Raman 2014) that were originally identified in cerebellar Purkinje neurons, which express Na_V1.6 (Raman and Bean 1997; Raman et al. 1997). Specifically, the intracellular tail of Na_Vβ4 has been implicated as the component responsible for resurgent currents (Grieco et al. 2005). Resurgent currents have also been observed in large DRG neurons expressing Na_V1.6 (Cummins et al. 2005). Thus our results would predict that channels exhibiting resurgent currents would be resistant to μO§-GVIIJ_SSC/G; conversely, channels that do not exhibit resurgent currents and yet are resistant to μO§-GVIIJ_SSC/G may be associated with Na_Vβ2. However, this assessment may be too simplistic in light of a recent report that a critical regulator of a resurgent current in cerebellar Purkinje neurons is intracellular FGF14 (Yan et al. 2014).

The block by saturating concentrations of μO§-GVIIJ_SSC/G of various Na_V1s expressed in oocytes is not 100%; i.e., there remains a “residual current” (Gajewiak et al. 2014). It is presently unknown how much of the incomplete block of the TTX-sensitive I_Na of small neurons by μO§-GVIIJ_SSG (Fig. 6) is due to channel heterogeneity and how much to partial efficacy. Heterogeneity in the oxidation state of the Cys residues of the channel can contribute to incomplete block, insofar as DTT treatment of Na_V1s expressed in oocytes reduces the residual current (Gajewiak et al. 2014); thus it would be interesting to see whether DTT treatment of DRG neurons or sciatic nerve can likewise increase the efficacy of block of endogenous channels by μO§-GVIIJ_SSC/G. In this regard, it might be noted that we have not determined the safety factor for the propagation of A- and C-CAPs under our experimental conditions; however, as mentioned in results in describing Fig. 4, μO§-GVIIJ_SSC/G produced no decrease in action potential conduction velocity; in other words, no evidence for partial block was detected.

Thus our current working hypothesis is that coexpression of Na_Vβ2 and/or Na_Vβ4 protects Na_V1s in both A- and C-fibers against block by μO§-GVIIJ. This proposition is concordant with other observations; that is, immunohistochemistry, PCR, and in situ hybridization indicate that all four Na_Vβ-subunit isoforms are expressed in DRG neurons (Coward et al. 2001; Takahashi et al. 2003; Yu et al. 2003) with Na_Vβ2 expressed in both large and small DRG neurons (Ho et al. 2012). Furthermore, the majority of the nodes of Ranvier in rat peripheral nerve is labeled by antibody against Na_Vβ4 (Buffington and Rasband 2013), and in view of our results, the nodes not expressing Na_Vβ4 presumably have VGSCs associated with Na_Vβ2 instead.

If we accept that Na_Vβ2 or -β4 is associated with Na_V1.6 in A-fibers, then it seems reasonable to suggest that Na_Vβ1 or -β3 is also part of the VGSC complex in view of the influence that binary coexpression of Na_Vβ-subunits has on percentage block of mNa_V1.6 by μ-SxIIIA. Block by 30 μM μ-SxIIIA of mNa_V1.6 coexpressed with β4 is expected to be only 73%, whereas 97% block is expected if the channel were coexpressed with β1+β4 or β3+β4 (Table 2). μ-SxIIIA, at concentrations of 30 μM, blocked mouse A-CAPs by ∼95% (Fig. 3I). In view of the safety factor for conduction of action potentials, a disproportionate fraction of sodium channels must be blocked relative to the reduction of the CAP; thus for μ-SxIIIA to be as potent as it is in blocking A-CAPs in mouse sciatic, Na_V1.6 in A-fibers is likely to be associated with β1 or β3 in addition to β2 or β4.

We assume here that the μ-conotoxin susceptibilities of VGSCs exogenously expressed in oocytes can be applied directly to channels endogenously expressed in neurons and axons, as we have done previously (Wilson et al. 2011a; Zhang et al. 2013b). μ-Conotoxins bind to site 1 near the selectivity filter and block the channel by interfering with Na⁺ conduction (Cestèle and Catterall 2000; Zhang et al. 2009, 2010a, b). On the other hand, μO-GVIIJ_SSC/SSG binds to site 8 between S5 and SS5 of the pore loop of DII, although its mechanism of block remains to be established (Gajewiak et al. 2014). Thus μ- and μO§-conotoxins (both of which are relatively hydrophilic) bind to extracellular aspects of the channel and may therefore be minimally influenced by the lipid composition of surrounding the plasma membrane. In this regard, it might be noted that the Na_V1-isoform selectivity of μO-GVIIJ_SSG is similar whether the channels are expressed in X. laevis oocytes or mammalian cell lines (human embryonic kidney 293 and Chinese hamster ovary cells) (Gajewiak et al. 2014).

Although it is most parsimonious to assume that coexpression of Na_Vβ2 or -β4 is responsible for protecting the endogenous VGSCs from block by μO§-GVIIJ_SSC/G, we cannot rule out that some other factor confers resistance of A- and C-CAPs to μO§-GVIIJ_SSC/G; e.g., the efficacy of the peptide in blocking endogenously expressed VGSCs that are not coexpressed with Na_Vβ2 or -β4 may be sufficiently small and/or the safety factor for propagation of action potentials sufficiently large that the inhibition or slowing down of action potentials by μO§-GVIIJ_SSC/G is not observable in CAPs. More sensitive means to monitor the effects of toxins on action potentials in the sciatic nerve are under investigation, such as experiments under reduced concentrations of extracellular Na⁺ or in the presence of low concentrations of TTX (Colquhoun and Ritchie 1972; Fern and Harrison 1993), as well as experiments to see whether μO§-GVIIJ_SSC/G alters the stimulus strength required to generate action potentials following placement of the toxin in the well containing the depolarizing stimulus electrode.

Thus far, we have examined only VGSCs on the cell bodies and bulk of the axons in the peripheral nervous system, and it would be interesting to examine whether μ- and μO§-conotoxins can be used to assess the likely molecular composition of VGSCs at nerve terminals, as well as at various locations of central nervous system neurons.

GRANTS

Support for this work was provided by the National Institute of General Medical Science (Grant GM 48677).

DISCLOSURES

The authors declare no competing financial interests.

AUTHOR CONTRIBUTIONS

Author contributions: M.J.W., M-M.Z., and D.Y. conception and design of research; M.J.W., M-M.Z., J.G., L.A., and J.E.R. performed experiments; M.J.W., M-M.Z., and J.G. analyzed data; M.J.W., M-M.Z., J.G., and D.Y. interpreted results of experiments; M.J.W., M-M.Z., J.G., and D.Y. prepared figures; M.J.W., M-M.Z., B.M.O., and D.Y. drafted manuscript; M.J.W., M-M.Z., J.G., L.A., B.M.O., and D.Y. edited and revised manuscript; D.Y. approved final version of manuscript.